Public health research: protocol writing and ICMR ethical guidelines

In short

A research protocol is the complete written plan of a study — the question, design, population, sample size, measurements, analysis and ethics safeguards — fixed *before* the first participant is enrolled, so that the study is reproducible and its findings are defensible. In India, biomedical and health research on human participants is governed by the ICMR National Ethical Guidelines for Biomedical and Health Research Involving Human Participants (2017), with drug and device trials additionally governed by the New Drugs and Clinical Trials Rules, 2019 (CDSCO) under the Drugs and Cosmetics Act, 1940. Every MPH dissertation, every ICMR/DHR/NHM-funded project and every NGO or industry study passes through an Institutional Ethics Committee (IEC); protocol writing and ethics are therefore examined in MPH research-methods papers, PhD coursework, and tested in interviews for research-coordinator, monitoring-and-evaluation, clinical-research-associate and programme-officer roles. Employers expect you to be able to draft a protocol, a participant information sheet and an informed consent form without supervision.

Explained

What a protocol is, and the standard sections

1. Title page — short descriptive title, version number and date (every amendment gets a new version), principal investigator (PI), co-investigators, institution, funder, protocol ID.
2. Abstract / synopsis — 250–300 words: background, objectives, design, setting, participants, sample size, outcomes, analysis.
3. Introduction and literature review — the public health burden with Indian numbers (NFHS-5, SRS, NCRB, GBD, NCDIR), the knowledge gap, and justification for doing the study in this population.
4. Objectives — one primary objective stated in measurable terms, a small number of secondary objectives; hypothesis for analytical studies. Avoid "to study" — use "to estimate", "to compare", "to determine the association".
5. Methodology — study design; study area and setting; study period; study population with explicit inclusion and exclusion criteria; sample size with assumptions and software; sampling technique; variables and operational definitions; data collection tools (validated scales, with permission to use); pilot testing and pre-testing; quality control.
6. Statistical analysis plan — software, descriptive statistics, tests for each objective, level of significance, handling of missing data and confounders, subgroup analyses declared in advance.
7. Ethical considerations — IEC approval, consent process, risks and benefits, confidentiality, data storage, compensation for research-related injury, conflict of interest.
8. Work plan and timeline — Gantt chart; budget with justification; references in a consistent style (Vancouver for biomedical journals).
9. Annexures — Participant Information Sheet (PIS), Informed Consent Form (ICF) in English and the local language, case record form (CRF), questionnaire, IEC letter, CTRI/registry details, PI's CV.

Framing the question and choosing the design

• PICOT — Population, Intervention/Exposure, Comparator, Outcome, Time. FINER test for feasibility: Feasible, Interesting, Novel, Ethical, Relevant.
• Observational: cross-sectional (prevalence, KAP surveys), case–control (odds ratio; good for rare outcomes), cohort (incidence, relative risk), ecological.
• Interventional: randomised controlled trial (individual or cluster RCT, common in community public health), quasi-experimental/before–after, stepped-wedge (useful when a programme must reach everyone eventually).
• Others: diagnostic accuracy (sensitivity/specificity, STARD), implementation research, health-economic evaluation, systematic review and meta-analysis (PROSPERO registration).
• Reporting checklists to write the protocol against: SPIRIT 2025 (trial protocols, updating SPIRIT 2013), CONSORT 2025 (trial reports), STROBE (observational), PRISMA/PRISMA-P (reviews), COREQ (qualitative), CHEERS (economic evaluations).
• Registration: interventional studies must be registered prospectively in the Clinical Trials Registry – India (CTRI), run by ICMR-NIMS; Indian journals and ICMJE require prospective registration for publication.

The four ethical principles and the twelve ICMR principles

• International basis: Nuremberg Code (1947), Declaration of Helsinki (WMA, 1964, latest revision 2024), Belmont Report (1979) — respect for persons, beneficence, justice — and CIOMS 2016 guidelines.
• Four principles (Beauchamp and Childress): autonomy, beneficence, non-maleficence, justice.
• ICMR 2017 lists twelve general principles, which exam answers should be able to name: (1) essentiality, (2) voluntariness, (3) non-exploitation, (4) social and financial responsibility, (5) maximisation of benefit, (6) institutional arrangements, (7) transparency and accountability, (8) totality of responsibility, (9) respect and protection of privacy and confidentiality, (10) risk minimisation, (11) professional competence, (12) compliance (with laws and guidelines).
• Risk is classified as less than minimal, minimal, minor increase over minimal, and more than minimal risk; the level decides the review route.

Ethics committee review in practice

• Composition of an IEC (ICMR 2017 / NDCT Rules 2019): minimum seven members (ICMR recommends 7–15) — a chairperson from outside the institution, a member secretary, one or two basic medical scientists, one or two clinicians, a legal expert, a social scientist or NGO/philosopher/ethicist representative, and a lay person from the community. At least one woman; not more than half from one profession; quorum must include the chair or secretary, a basic scientist, a clinician, a lay person/social scientist, and at least one non-affiliated member.
• Registration: an ethics committee reviewing regulated clinical trials must be registered with CDSCO (Form CT-01, registration normally for five years and renewable); committees reviewing only academic biomedical and health research register with the DHR/National Ethics Committee Registry for Biomedical and Health Research (NECRBHR).
• Review routes: (a) exemption — e.g. anonymised publicly available data, routine educational practice, programme audit (the IEC, not the investigator, decides exemption); (b) expedited/expert review — minimal risk, minor amendments, re-review in an emergency; (c) full-committee review — everything else, especially vulnerable populations and more-than-minimal risk.
• Decisions: approve, approve with recommendations/revision, defer for clarification, reject. Approval is time-limited — annual (or more frequent) progress reports, prior approval for amendments, reporting of serious adverse events (SAEs), and a final report on completion.
• Serious adverse event reporting in a regulated trial: investigator to the ethics committee, sponsor and licensing authority within 24 hours, followed by a detailed analysis by the investigator, IEC and sponsor within 14 days of the first report; compensation for a trial-related injury or death is paid by the sponsor as per the NDCT Rules 2019 formula.
• A protocol may not be amended, nor data collected, before written IEC approval; retrospective ("post-facto") approval is not acceptable and journals reject such work.

Informed consent, privacy and data

• Elements of valid consent: disclosure of purpose, procedures, duration, risks, benefits, alternatives; statement that participation is voluntary and withdrawal carries no penalty; confidentiality arrangements; compensation and treatment for injury; contact details of PI and IEC; language the participant understands; adequate time to decide; no coercion or undue inducement.
• Documentation: signed/thumb-impression ICF plus a copy of the PIS given to the participant; impartial witness when the participant or their representative cannot read; audio-visual consent is required for vulnerable participants and for new-chemical-entity and HIV/leprosy-related trials under Indian rules; LAR (legally acceptable representative) consent plus the child's assent (ICMR: assent from children aged roughly 7–17; re-consent at 18).
• Waiver of consent may be granted by the IEC only when research could not practicably be done otherwise, risk is minimal, the waiver does not adversely affect rights and welfare, and the research is of public benefit — typical in retrospective record reviews, outbreak investigations, and research on anonymised stored samples.
• Vulnerable groups needing extra safeguards: children, pregnant and lactating women, prisoners, persons with mental illness or cognitive impairment, economically or socially disadvantaged persons, tribal communities, students and employees of the investigator, terminally ill patients, sex workers and other stigmatised groups.
• Community consent / gatekeeper permission (village panchayat, tribal council, school principal) is additional to — never a substitute for — individual consent.
• Privacy law: the Digital Personal Data Protection Act, 2023 governs digital personal data; its DPDP Rules, 2025 phase the Data Fiduciary duties (notice, consent, breach intimation, erasure) and the omission of IT Act section 43A so that they apply only from 13 May 2027 — until then section 43A of the IT Act, 2000 and the SPDI Rules, 2011 apply to sensitive personal data. Research use is exempt under section 17(2)(b) of the DPDP Act where the processing is carried on in accordance with the standards prescribed in the Second Schedule to the DPDP Rules, 2025.
• Confidentiality in practice: coded identifiers, a separate password-protected linking key, restricted access, retention for the period the IEC specifies (institutional norms commonly 3–5 years after publication, longer for trials), and specified destruction.

Specific Indian regulatory and governance requirements

• Biological material and biobanking: separate consent for storage and future use; Material Transfer Agreement (MTA) for sharing samples; export of human biological material requires clearance under the Health Ministry's Screening Committee (HMSC) route for international collaboration.
• International collaboration / foreign funding: HMSC approval for collaborative health research, and compliance with the FCRA, 2010 for foreign contributions.
• Genetic and biodiversity aspects: Biological Diversity Act, 2002 (as amended 2023) and the Nagoya Protocol for access and benefit-sharing where biological resources or traditional knowledge are used; ICMR guidelines on genetic testing and gene therapy (the ICMR National Guidelines for Gene Therapy Product Development, 2019 with CDSCO).
• Stem cells: ICMR–DBT National Guidelines for Stem Cell Research (2017), with an Institutional Committee for Stem Cell Research (IC-SCR).
• ART and surrogacy research: Assisted Reproductive Technology (Regulation) Act, 2021 and Surrogacy (Regulation) Act, 2021.
• Sex determination: the PCPNDT Act, 1994 restricts disclosure of fetal sex in any research involving prenatal diagnostics.
• Animal work (if any component): CPCSEA rules and the Institutional Animal Ethics Committee; now administered under the Prevention of Cruelty to Animals framework.
• Scientific misconduct: fabrication, falsification, plagiarism (FFP), duplicate publication, salami slicing, gift/ghost authorship. UGC (Promotion of Academic Integrity and Prevention of Plagiarism) Regulations, 2018 prescribe levels of similarity and penalties; ICMJE four criteria govern authorship; COPE flowcharts govern retractions.
• Good practice standards: ICH-GCP E6, Indian GCP guidelines (CDSCO, 2001), ISO standards for biobanks, and a written SOP for every ethics committee.

Sample size, budget and feasibility — the parts protocols fail on

• Every sample size must state: the parameter assumed (prevalence, mean, difference), its source with citation, confidence level, power, design effect for cluster sampling, non-response inflation, and the software or formula used.
• Budget heads typically allowed by ICMR/DHR extramural grants: manpower (JRF/field staff), consumables, travel, equipment, contingency, and institutional overhead; honorarium/reimbursement to participants is for time and travel only and must not be large enough to induce participation.
• Feasibility check before submission: permissions (District Medical Officer, CMO, school, factory), translation and back-translation of tools, trained field staff, pilot on roughly 10% of the sample (pilot data excluded from the main analysis), and a realistic timeline.

Key points

• Protocol — the complete pre-specified plan; version-controlled, with every change approved as an amendment.
• Primary objective — exactly one, measurable; it drives the sample size and the primary outcome.
• ICMR 2017 guidelines — India's core ethics document for biomedical and health research on humans, replacing the 2006 edition.
• NDCT Rules, 2019 — regulate clinical trials, bioavailability/bioequivalence studies and ethics committee registration for regulated trials.
• CTRI — prospective registration of interventional studies, mandatory for publication in ICMJE-compliant journals.
• IEC composition — at least seven members including a lay person, a legal expert, a social scientist and a non-affiliated chair; at least one woman.
• Review routes — exemption, expedited and full-committee; only the IEC may declare a study exempt.
• Four principles — autonomy, beneficence, non-maleficence, justice; ICMR expands these into twelve principles including essentiality and non-exploitation.
• Risk categories — less than minimal, minimal, minor increase over minimal, more than minimal.
• Informed consent — a process, not a signature; PIS plus ICF in the local language, with the right to withdraw.
• Assent — a child's own agreement (roughly 7–17 years) alongside LAR consent; re-consent on turning 18.
• Waiver of consent — permitted only for minimal-risk research that is impracticable otherwise and of public benefit.
• SAE reporting — within 24 hours to IEC, sponsor and licensing authority; detailed analysis within 14 days in a regulated trial.
• Vulnerable populations — children, pregnant women, prisoners, tribal groups, mentally ill, economically disadvantaged; need extra safeguards, not exclusion by default.
• Reporting checklists — SPIRIT (protocol), CONSORT (RCT), STROBE (observational), PRISMA (reviews), COREQ (qualitative).
• HMSC and FCRA — clearances for international collaborative health research and foreign funds.
• Privacy — DPDP Act, 2023 with Rules, 2025 (Data Fiduciary duties from 13 May 2027); until then IT Act s.43A and SPDI Rules, 2011 apply.
• Misconduct — FFP, plagiarism (UGC Regulations, 2018), ICMJE authorship criteria, COPE retraction flowcharts.

Formulas and examples

Example 1 — prevalence survey sample size (anaemia among adolescent girls). Expected prevalence p = 0.59 (NFHS-5 for adolescent girls, taken as 59%), absolute precision d = 0.05, 95% confidence.

• p(1−p) = 0.59 × 0.41 = 0.2419.
• n = (1.96)² × 0.2419 / (0.05)² = 3.8416 × 0.2419 / 0.0025 = 0.92928 / 0.0025 = 371.7 → 372.
• Add 10% non-response: 372 / 0.90 = 413.3 → 414 girls.

Example 2 — the same survey by 30 clusters of 14 girls each. ICC ρ = 0.05, cluster size m = 14.

• DE = 1 + (14 − 1) × 0.05 = 1 + 0.65 = 1.65.
• n = 372 × 1.65 = 613.8 → 614; with 10% non-response, 614/0.90 = 682.2 → 683, i.e. about 23 girls from each of 30 clusters (690). The protocol must show this inflation — omitting the design effect under-powers the study by 40%.

Example 3 — two-arm cluster-randomised trial of a handwashing intervention on diarrhoea. p₁ = 0.20 (control), p₂ = 0.12 (intervention), α = 0.05 two-sided, power 80% (Z = 1.96, 0.84).

• p̄ = 0.16, q̄ = 0.84 → 2p̄q̄ = 0.2688, √0.2688 = 0.5185; 1.96 × 0.5185 = 1.0162.
• p₁q₁ = 0.20 × 0.80 = 0.16; p₂q₂ = 0.12 × 0.88 = 0.1056; sum = 0.2656, √ = 0.5154; 0.84 × 0.5154 = 0.4329.
• Numerator = (1.0162 + 0.4329)² = (1.4491)² = 2.0999. Denominator = (0.08)² = 0.0064.
• n = 2.0999 / 0.0064 = 328.1 → 329 children per arm; with DE = 1.65, 543 per arm, 1,086 in all.

Example 4 — budget for the survey in Example 1 (414 participants).

Example 5 — choosing the review route. Three protocols reach an IEC: (a) secondary analysis of fully anonymised NFHS-5 data — exempt from review, but the exemption is recorded by the IEC, not assumed by the student; (b) a KAP questionnaire among healthy adults with no identifiers beyond a code — minimal risk, expedited review; (c) finger-prick haemoglobin testing in 414 adolescent girls — a minor increase over minimal risk involving minors, so full-committee review, LAR consent plus assent, and a referral pathway for girls found severely anaemic (a benefit-maximisation duty, not optional).

Where it is used

• Academic and government research — MPH/MD dissertations, ICMR and DHR extramural grants, AIIMS/PGIMER/NIE/NIRT/NIN institutional projects, State NHM operational research.
• Clinical research industry — CROs and pharma (Syngene, Lambda, Veeda, Sun, Cipla, Dr Reddy's, Serum Institute) employ clinical research associates and coordinators to write, submit and monitor protocols under the NDCT Rules.
• NGOs and development agencies — PHFI, St John's Research Institute, SNEHA, Jhpiego, PATH, Ikp Centre for Technologies in Public Health; baseline/endline surveys and impact evaluations with IEC approval.
• Programme monitoring and evaluation — M&E officers designing LQAS and concurrent evaluations for NHM, NTEP, NCVBDC (formerly NVBDCP), and Poshan Abhiyaan; here the ethics boundary between programme audit and research matters.
• Ethics committee roles — member secretary and IEC coordinator posts in medical colleges and hospitals; NABH accreditation and CDSCO inspections check the SOP, minutes and quorum records.
• Regulatory and consulting — regulatory affairs roles preparing CDSCO submissions; insurers and sponsors assessing trial-injury compensation claims.

Interview questions

• Q: What is the difference between a research protocol and a research proposal? A: A proposal is written to persuade a funder or committee and includes budget, justification and investigator credentials; the protocol is the technical plan of exactly how the study will be conducted, and is the document the ethics committee approves and the team follows.
• Q: Which document governs health research ethics in India, and which is the statutory one? A: The ICMR National Ethical Guidelines, 2017 are the national ethical standard; the statutory instrument for clinical trials is the New Drugs and Clinical Trials Rules, 2019 under the Drugs and Cosmetics Act, 1940.
• Q: Who decides whether a study is exempt from ethics review? A: The ethics committee, never the investigator — the investigator must still submit and obtain a written exemption.
• Q: What is the minimum composition of an Institutional Ethics Committee? A: At least seven members including a non-affiliated chairperson, member secretary, basic medical scientist, clinician, legal expert, social scientist or ethicist, and a lay person from the community, with at least one woman and no profession in a majority.
• Q: When can informed consent be waived? A: When the research is minimal risk, could not practicably be carried out with consent, the waiver does not adversely affect participants' rights and welfare, and the research serves public benefit — the IEC must grant it explicitly.
• Q: Difference between consent and assent? A: Consent is legally valid permission given by a competent adult or the legally acceptable representative; assent is the affirmative agreement of a child or person with limited capacity, which must also be sought and respected.
• Q: You need to publish an RCT but forgot to register it. What happens? A: ICMJE journals and the CTRI require prospective registration, so retrospective registration usually leads to rejection or a disclosed deviation — the fix is prevention, registering before enrolling the first participant.
• Q: How soon must a serious adverse event be reported in a regulated trial in India? A: Within 24 hours to the ethics committee, sponsor and licensing authority, with a detailed analysis within 14 days of the first report.
• Q: What is a design effect and why does it change your sample size? A: DE = 1 + (m − 1)ρ accounts for people in the same cluster being more alike than people across clusters; it inflates the simple random sample size, so ignoring it under-powers a cluster survey.
• Q: A sarpanch agrees on behalf of the whole village. Is that enough? A: No — community or gatekeeper permission is an additional requirement, and individual informed consent from each participant is still mandatory.
• Q: What are the ICMJE criteria for authorship? A: Substantial contribution to conception/design or data acquisition/analysis; drafting or critical revision; approval of the final version; and accountability for the work — all four, which is why gift and ghost authorship are misconduct.
• Q: What privacy law applies to a digital health survey in India today? A: The DPDP Act, 2023 is enacted, but under the DPDP Rules, 2025 the Data Fiduciary duties begin on 13 May 2027; until then section 43A of the IT Act, 2000 and the SPDI Rules, 2011 govern sensitive personal data, alongside the ICMR confidentiality requirements.

Common mistakes

• Writing many "primary" objectives, then calculating sample size for something that is not the primary outcome.
• Stating a sample size without the assumed parameter, its source, precision, power, design effect or non-response allowance — the commonest reason a protocol is deferred.
• Starting data collection, or amending the method, before written IEC approval, and then seeking post-facto approval.
• Treating informed consent as a signature on a form rather than a documented process of disclosure, comprehension and voluntariness in the participant's own language.
• Excluding women, children or tribal participants "to be safe" — unjustified exclusion is itself an injustice; the requirement is extra safeguards, not blanket exclusion.
• Confusing anonymised with coded data: if a linking key exists, the data are identifiable and need full confidentiality safeguards.
• Mislabelling research as programme evaluation or audit to avoid ethics review, and omitting prospective CTRI registration for an interventional study.

Related topics

Qualitative research methods in public health · Study designs and epidemiological measures · Sampling techniques and sample size estimation · Biostatistics: hypothesis testing and confidence intervals · Health policy, legislation and regulatory bodies in India · Scientific writing, systematic reviews and publication ethics